The launch of Alphram has resulted in many posts regarding its relationship to Google, and its ability to deliver whatever it is meant to deliver. Certainly, I see it as being a different beast from traditional search, and I regard the initial period of its release as a time in which users will be looking for the systems natural niche. This search will basically be the search for the intersection of expectations and execution.
In our discussions last night following the opening tutorials for ICWSM the topic of the carbon footprint of our conference came up. It occurred to me that this would be a perfect Alphram style query – what is the carbon cost of a flight from Seattle to San Jose? Better yet, compare this with the same journey by train. Given my understanding of the system and having viewed the screencast, my expectation was that this would be where the system could deliver.
Unfortunately, no such luck. I can get it to say something about the distance (seattle san jose) and about the time via plane (seattle san jose airplane), about carbon emissions in general (carbon emissions). Seattle carbon reveals that there is a town in Mexico called carbon (why would that be preferred to the chemical interpretation?).
Is this an extensibility issue (how will Alphram extend its capabilities to estimate some derived cost of a quantity – e.g. the carbon emissions of a flight of a certain duration)? Or a query interpretation problem (it can do it, but I can’t get the query right)?